Imaging projects
Browser visualisation projects of self-developed software or implementations of published third-party research.
Groundwork
The projects below draw on the same teaching, collected in the study track: 394 lecture courses, 335 of which play in the page. Each project names the lectures it rests on.
Medical imaging · Computer vision · Deep learning · Optimization · General machine learning
-
AI Imaging/Segmentation
-
Beta
DICOM viewer
Read and window DICOM studies in the browser.
- Schnabel, Medical imaging meets deep learningWhat the modalities are and what they measure.
- Rueckert, Segmentation and reconstructionWhat is done with a volume once it is loaded.
-
-
3D reconstruction
-
Beta
Find the C-arm position
DRR pose recovery in the browser.
- Multiple View Geometry, lecture 2Rigid-body motion: the six numbers being solved for.
- Multiple View Geometry, lecture 3Perspective projection, which is what a DRR imitates.
- Wells, Uncertainty in registrationRegistration posed as an optimisation.
-
In progress
Cinematic rendering for 3D CT
Photorealistic CT volumes.
The study track has no volume rendering or light transport course, so this one still needs a graphics reference from outside it.
-
-
Machine Learning
-
In progress
aleArrhythmia
Cardiac arrhythmia mapping.
- MIT 6.S897, physiological time seriesLearning from a signal rather than an image.
- MIT 6.S897, machine learning in cardiac imagingThe clinical end of the same problem.
-
In progress
CT reconstruction with machine learning
Learned image reconstruction.
- Hinton, Neural Networks, lecture 1.1The fundamentals, if you are starting cold.
- Rueckert, Deep learning for image reconstructionThe method this project implements.
- Rueckert, Segmentation and reconstructionThe earlier, classical framing.
-